OMQ Update | Version Lima 10.2 - Self Service Automator and GPT improvements

We have released the product version Lima 10.2.

With version 10.2 Lima it is now possible to create simple action automators independently. Other highlights include GPT improvements for the Automator and conditions and the automatic extraction of information from messages (e.g. email addresses).

To better understand the new features, in this article we show various use cases such as automated contact data capture, improved complaint processes and redirects to product pages.

Create simple action automator independently

What is an Automator?

Questions and answers are usually static, so the question: “Where is my order?” has a specific, predefined answer. Automators, on the other hand, are dynamic and contain response actions that are interactive.

For example, if customers want to change their contact details, an automator can respond by requesting data, receiving it from the customer and then processing it independently in the system.

How do you create an Automator?

With the Lima 10.2 version update, simple action automators can now be created independently. The new feature enables more autonomy in the automation of customer service by creating interactive response actions in the editor without IT knowledge.

Various elements can be inserted to create and individualize it and to query customer data:

• Input field
• Selection field
• Upload
• Date selection

In the second step, commands can be used to determine which actions are to be executed. On the one hand, there are chatbot commands:

• End chat
• Start chat
• Navigate to website

On the other hand, live chat commands:

• Update contact information
• Update conversation information
• Forwarding by e-mail or agent forwarding, e.g. to groups
• Add notes
• Edit subject

If you need other automators that, for example, change the order status or query other information such as data from Excel tables from backend systems, you can contact us.

1. Example use case: Query contact data

In our first application example, a contact data query field is created in the new “Automator” module, which queries the customer’s name and email address, among other things.

The components to be queried are defined in the input and then used in the output tab. Curly brackets are used for this (e.g.: {{email}}). This ultimately fills in the corresponding information.

This form can be displayed in the chat, whereby the contact details are requested by the chatbot and then changed in the backend of the live chat system.

This eliminates the need for manual changes by agents, saving time and effort by optimizing processes such as data retrieval.

2. Example use case: Navigating to the product page

The Simple Action Automator recognizes intentions in customers’ messages and then provides them with the desired information accordingly. If a customer asks about a specific product, the Automator can take them directly from the chat to the product page.

Information about the product is stored in the database and customers can then be forwarded directly to the product page.

Additional commands can also be stored in the Automator, for example the option to send a message to the operator groups that the customer is interested in an item. This information can then also be noted in the backend using the note commands and the subject assignment of the conversation.

3. Example use case: Forwarding the chat conversation

In this use case, information on damaged products is requested in the chat and then forwarded by the Automator, with the conversation being sent in the form of a ticket either by email or directly to an operator in the live chat system. This eliminates the need for time-consuming writing back and forth.

By creating and inserting conditions, helpful hints for solving the problem can be given during the complaint process.

The selection of forwarding functions allows great flexibility in the further processing of the complaint, which guarantees seamless transitions and faster responses, as the service agents already receive all the information in advance. In addition, elements such as file upload can be used to obtain detailed descriptions from customers.

Forwarding options include e-mail and ticket forwarding (e.g. to Zendesk, Salesforce or other ticket systems).

GPT Bot with interactive elements

If customers describe several concerns in one message, the GPT Bot recognizes the individual questions. As the GPT Bot is equipped with interactive elements such as the Automator and conditions, it can reliably change information in the backend of the live chat system.

For example, if the customer wants to change both their email and delivery address, the GPT Bot gives them the opportunity to make these changes via Clickbot.

This results in a continuous dialog flow, as the customer does not have to send several separate messages and the number of queries is reduced.

Automatic extraction of information from messages

The automator immediately recognizes names and data in messages. Previously, several messages or requests from the bot were necessary for this.

Now, one message from the customer containing all the information is sufficient. Multiple requests are no longer necessary and fewer messages are needed to collect all the required information.

Test our Automator and GPT technology

If you would like to try out the ChatGPT chatbot and our Automator, you can request a demo without obligation or simply contact us.

Further new changes:

• Fallback options for conditions
• Translate Glossary

OMQ Help

• Inline styles are now displayed in responses.

• Shortcut elements have been implemented for OMQ Help.

• The subcategories of the FAQ categories are now displayed in alphabetical order.

• For the Help connections, incorrect languages were sometimes displayed for the FAQ categories. This problem has been fixed.

• Search results can now also be displayed inline.

• A special message has been added that is displayed when a category contains no entries.

OMQ Chatbot

• Forwarding to employees is now possible using skills.

• The connection to Userlike has been expanded to include an asynchronous variant.

• Inline styles are now displayed in message bubbles.

• Fixed an issue where non-text-based input caused a problem.

• The use of Userlike environment variables is now possible.

• The Userlike chatbot can now also be accessed via a different URL. This has optimized access.

• Fallback options are now displayed for conditions if the user’s input does not meet the possible inputs.

• If there is a timeout when processing an input, an error message is now displayed.

• Formatted text was not displayed correctly in some cases. This problem has been fixed.

• The security mechanism has been adapted for the Userlike chat interfaces.

• The display that the chatbot is currently doing something has been improved.

• Videos can now be displayed by the GPT chatbot.

• Fixed a problem when processing some WhatsApp messages.

• Some conditions were not processed due to case sensitivity. This issue has been fixed.

• Click inputs were sometimes processed incorrectly. This issue has been fixed.

• Userlike commands have been added.

• Feedback is now a default setting.

• A problem with the formatting of messages has been fixed.

• Some points were displayed that should not be displayed. This problem has been fixed.

OMQ Contact

• Inline styles are now displayed in responses.

• Shortcut elements have been implemented for OMQ Contact.

• Feedback is now a default setting.

OMQ Assist

• Inline styles are now displayed in responses.

OMQ Reply

• Inline styles are now displayed in responses.

• A function has been implemented that regularly collects emails from the server.

• The retrieval of emails has been distributed across several application servers.

• The parallel processing of e-mail responses has been reduced.

OMQ Automator

• The e-mail automator can now also be connected to external e-mail servers.

• In some cases, e-mails could not be sent due to file uploads. This problem has been fixed.

• The Automator has been completely rebuilt.

• The properties have been adapted and improved in several steps.

• The user interface has been adapted and revised in several steps.

OMQ Engine

• An adapted model for the tourism sector was trained.

• The model for tourism has been improved.

• The recognition of small talk dialogue has been improved.

• Logging for the GPT Bot has been improved.

• Fixed an issue where the GPT Bot could not display responses with conditions.

• The processing of Automator and conditions has been improved.

• A newer GPT model has been connected.

• Messages about filtering LLM are now processed and displayed in the user output.

• The index server is now updated in even more cases than newly created.

• The connection to GPT has been improved.

• The GPT pipeline has been revised.

• Inline links are now displayed correctly in the GPT bot.

• Word vectors are now processed differently. The size is checked.

• Fixed a problem where phone numbers were created.

• The response of the GPT interfaces has been improved.

• The connection of button links has been improved.

• Entities can now be extracted from chat messages.

• The connection to the index server has been improved.

• The GPT Bot now responds better to general dialogs.

• In some cases, the conditions were not played. This problem has been fixed.

• Forwarding to an agent has been improved for the GPT Bot.

• In some cases, the index was completely rebuilt, although this was not necessary. This problem has been fixed.

Administration

• Elements could not be deleted in the Safari browser. This problem has been fixed.
• In some cases, entries in the database could not be saved. This problem has been fixed.
• Problems with editing knowledge base articles have been fixed.
• Searching for knowledge base entries on the Safari browser was incorrect. This problem has been fixed.
• Incorrect passages of knowledge base entries are now marked with a warning instead of an error.
• Various incorrect labels have been improved.
• The UI is now prevented from being cached with a version code. In the past, this caused the old version to be incorrectly updated.
• A variant of the condition that is used as the default option has been implemented.
• Fixed a problem where the save button could be pressed twice in quick succession.
• Fixed a problem where the message that the entry is being trained was not displayed.
• Loading categories with many entries was very slow. This problem has been fixed.
• It could happen that some links in the knowledge base were saved incorrectly. This problem has been fixed.
• The fallback option has been given settings.
• “Created by” and “Modified by” were not displayed correctly.
• Incorrect links made it impossible to edit entries. This issue has been fixed.
• A title has been added for video elements.
• The sorting display could not be closed in some cases. This problem has been fixed.
• Several help elements have been introduced in the administration.
• Titles can now be set for replies.
• Loom videos can now be linked.
• Share links can now be used for embedding videos.
• In some links, the languages of the categories were displayed incorrectly. This problem has been fixed.
• The login for the Partner Dashboard has been fixed.
• In some cases, the number of entries in a category was displayed incorrectly. This issue has been fixed.
• The Automator page has been completely revised and released for users.
• The following commands can now be used: Restart chat, End chat, Navigate to a page, Connect the user to a coworker, Send the chat to an application or to an e-mail address, Update the information of a conversation, Update the information about the user, Add a note.
• Problems can occur when searching for categories. This problem has been fixed.
• The interactive elements in the editor have been revised.
• Api keys can no longer be changed.
• Properties for the GPT chatbot can now be set on their own.
• A glossary function for translations has been added.
• Tooltips and labels have been revised.
• Fixed an issue where the Automator could not be deleted.

General

• The index server has been updated.

• More resources have been made available for the index server.

• A test for database evolutions has been created.

• The web server framework has been updated.

• The backend has been converted to a new language version.

• The connection of partner applications has been improved.

• The test server has been optimized.

• The output at the interfaces has been reduced. Unnecessary information is no longer displayed.